How to serialize an array in python using pickle?

Question

Hello I have a problem when I tried to serialize an array with python, I have the dictionary

d={}
d['data']= [array([2, 5, 3], dtype=uint8), array([4, 1, 9], dtype=uint8)]

but my error is:

'array' is not defined

so, I have 2 questions:

1.- how can I do an array like:

array([2, 5, 3],[4, 1, 9], dtype=uint8)

to create the array I use

append()

and how can I serialize the dictionary with the form of the array(question 1) using the next function

with open('outp', 'wb') as fo:
   pickle.dump(d, fo, 2)

If I remove the words 'array' and 'dtype=uint8' it has no errors, but I think that is not the correct form to serialize an array Thanks

from numpy import array

donkopotamus
– donkopotamus

2016-11-08 02:18:02 +00:00
Commented Nov 8, 2016 at 2:18 — donkopotamus
– donkopotamus, Commented Nov 8, 2016 at 2:18

ted · Accepted Answer · 2016-11-08 02:37:30Z

If I remove the words 'array' and 'dtype=uint8' it has no errors

This is beacause without these words then you are simply creating a list which can be dumped by pickle indeed.

However the array you are mentionning seems like a numpy array, you can simply add from numpy import arrayin the beginning of your script.

Also, you can not make this array([2, 5, 3],[4, 1, 9]) because a 2 dimensional array with numpy is created like this : array([[2, 5, 3],[4, 1, 9]])

then once you have your array you can dump it like this:

import pickle
import numpy as np
d={}
d['data']= [np.array([2, 5, 3], dtype=np.int8), np.array([4, 1, 9], dtype=np.int8)]
with open('outp', 'wb') as fo:
    pickle.dump(d, fo, 2)

Community · Accepted Answer · 2017-05-23 10:32:37Z

0

1) According to the Numpy docs, there is a builtin method for that. Check ndarray.dump.

2) According to this entry you can wrap the dictionary in an object and save it.

edited May 23, 2017 at 10:32

CommunityBot

11 silver badge

answered Nov 8, 2016 at 2:37

zenlc2000

4514 silver badges10 bronze badges

Collectives™ on Stack Overflow

How to serialize an array in python using pickle?

2 Answers 2

Comments

Comments

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Linked

Related